Last Updated: September 29th 2026
Note: The opinions in this article are from the viewpoint of someone in Human Computer Interaction fields. Other fields may alreay have this in place.
I recently completed a research study and submitted it for publication at a peer-reviewed conference. After a few rounds of review, it got rejected. This isn’t atypical in a scientist’s life: you submit papers, sometimes they get accepted, sometimes they do not. However, the reason for rejection really struck a nerve. It wasn’t rejected because it was bad science. It wasn’t rejected because my methodology was flawed, my experiments were biased, my data was misleading, or anything like that. In fact, all reviewers agreed that the science and methodology were top notch.
Instead, complains centered around abstract themes: the paper was not relevant enough, the contributions were confusing, the story was unclear, the reviewers did not see the expected impact, the problem was not well-motivated enough, and other such issues.
In other words, I had produced good knowledge and good data, but people did not really think it was a worthwhile problem to solve.
It was not the first time one of my papers, and for those in the HCI world, I bet you have faced a version of this comment against your work in some way or another. In human computer interaction, many papers are rejected for these reasons, and as a result, we dedicate a lot of time towards motivating the problem, articulating the contribution, and more. When I write a paper, i would say about 35% of the time is spent writing up the study design and the findings. The vast majority of time spent on a paper is spent either in the introduction (how do I present the problem in a way that makes it seem like the most existential question in the history of mankind?), and in the discussion section (how do I convince the reviewers that my findings are the best thing since sliced bread?).
I thought this was how all of science was. You spend some time actually studying a question, and some more time (sometimes even the majority of time) hyping up both the problem you are solving and the expected impact of your findings.
But then I read a neuroscience paper. Its title eludes me, but it had the shortest introduction I had ever seen – something along the lines of:
Many scientists have studied how fruit flies react to moving objects. But no one has yet studied how flies react to spherical objects moving downhill. In this paper, we carry out an experiment to answer this question.
That was it. No long convoluted paragraphs about why it is of utmost importance to understand how flies react to spherical objects moving downhill, no trying to convince reviewers that this was a worthy cause, no self-aggrandizement to the point of parody – none of that. Just here is a question that is yet unanswered, we tried to answer it followed by a multi-page in-depth look at the experiments they ran and their results.
It blew my mind. It was the simplest, most elegant paper I had read. I did not need to get bogged down in the life story of the authors, the history of the world, or anything like that. It was just: here is a problem, here is how we solved it.
This is what lead to my hot take.
HCI should stop rejecting papers based on relevance, expected contribution, or storyline. Instead, we should prioritize and judge papers primarily on rigor.
I know, fellow HCI folks, this is a controversial statement. Bare with me as I lay out my argument. I will go through three stages:
The status quo, where a large emphasis is based on storyline, perceived contribution, and relevance leads to several problems in the way that science is conducted.
Papers cannot properly explain their methods or the full extent of their findings. This problem stems from the fact that many venues in our field have limits to how long a paper is, whether through word count or paper length. Inevitably, this means that writing becomes a zero-sum game in that the more words and pages you spend on one part of the paper, the less you can spend on the other. Ergo: if you spend lots of words and pages fleshing out the introduction and the discussion to make sure the story fits, the less words and pages you have to describe what you did and your findings. This means people may not be able to assess the methods you used (either to replicate your study or assess the credibility of your claims) and some important findings may be omitted.
This encourages scientific dishonesty. One problem with emphasizing contribution, storyline, and relevance is that surprising, weird, and amazing findings make a better story than boring, mundane findings and negative results. In other words: impressive findings (we dove into Meta’s privacy policy and found that Mark Zuckerberg is stealing your data and selling it to Palantir henchmen so they can gamble it away with scientologists!) are more likely to be published than non-impressive findings (we went through Meta’s privacy policy and it was kind of boring.) So if impressive findings are more likely to get published, and a scientists’ career depends on getting published, where do you think the incentives are? To tell the truth? Or to embellish it to increase the chances of the paper getting in? But Abraham, you may be saying, no one would deliberately lie on their papers. Sure, outright fraud might be a step too far for many. But there are lots of smaller tricks people can do:
We mortgage our future by focusing on the present A major issue with relevance and importance is that it is incredibly short-sighted. The reviewers who appraise a study can only really imagine relevance and contribution for the here and now. Historically, that has not been the usefulness of science. Liebniz invented the binary code system in the 1700s. Did he have an idea of how relevant this would be for computing, and that his system underpins all of modern infrastructure or life? Probably not. Many modern encryption and communication algorithims are based on math that was done decades and possibly centuries earlier; and the people who invented that probably had no idea of their use. Where would we be if back in the day of Leibniz, Newton, and the like, their results had not been published because there was no immediate impact of their papers on the yield of potato crops?
Per Richard Feynman, “Physics is like sex: sure, it may give some practical results, but that’s not why we do it.”.
Wasted time This is a bit of a personal rant. Following the paper rejection outlined above, I spent the next few weeks rewriting and re-rewriting the paper to try to make it seem ‘cooler’ and more relevant. Did I advance science in any meaningful way? Did I discover any new data, probe new research directions, teach young minds? No, I wasted that time convincing three random people that these findings matter.
This does not only apply to this instance, but more broadly. Be honest. How much time and energy do you spend making sure the introduction and discussion section are just right? And isn’t that time, effort, and energy that could be better spent on doing actual science and gathering new and meaningful data?
Instead, imagine the world I am inviting you into. One where the storyline and contribution only matter a minimal amount. Instead, the primary focus of a paper’s strength is based on its methodological rigor. Do the methods make sense given the research question the paper is trying to answer? Were the methods conducted well, or conducted sloppily?
I am not arguing completely removing any form of relevance or contribution. A paper needs to be somewhat related to the venue it is being submitted to (an HCI venue should not review a treatise on the impact of roman mythology on medieval england’s Magna Carta). And it has to have some findings. And it has to be a problem that hasn’t already been answered 190548743257 times before (or it can, but then it has to clearly be labelled as a replication study). But it is vastly minimized, and barring a paper that goes completely off the rails, it is not a reason for rejecting a paper. Introductions are 1-2 paragraphs at most, discussion sections focus primarily on explaining anomalies or unexplained issues in the findings and/or how the findings differ from prior work (if at all) – importantly, a paper does not get rejected just because it’s results are similar to prior work. Authors can have their ‘Implications for design’ section should they wish, but they are not required to.
This would have the following benefits:
Wouldn’t this be ideal?
Here I walk through counterarguments to my claim and then rebut them. These counterarguments are a mixture of thoughts I had myself as well as critiques from colleagues.
Abraham you fool. Your approach would lead to a lot of papers with zero meaningful contributions being published Contribution is in the eye of the beholder. There is no thing that is ‘objectively’ a big or a small contribution. Clearly at least one person thought the paper had a meaningful contribution (i.e., the person who conducted the study in the first place). Someone else could think the contribution is important. Thus, I do not think it is wise to block what a minority of people would consider important based on what the majority would consider to be not important.
Furthermore, this ties back in to the issue I mentioned before that we are kneecapping our future by focusing on short-termism. A paper might have a seemingly minor contribution in 2026 but whose importance may only become apparent in 2090. I do not see publication of papers that make a minor contribution as a problem.
Finally, there is already a mechanism for prioritizing papers with more meaningful contributions: the grant application process. Studies often get funded by grants. When applying for a grant, a committee evaluates the application and judges it based on, among other factors, the expected contribution of the project to the field and to society. Thus there is already at some level a basic level of prioritization for projects that have more contributions. Adding it to the paper process as well as the grant application process is overkill in my opinion. If the work is done well, it should be allowed out there.
Academia already suffers from excessive publishing of papers. Without being able to reject papers based on relevance, the floodgates will open and we will be drowned in useless slop This is a legitimate concern, but then I have a solution. Start raising the bar extremely high for rigor. Don’t gatekeep papers on whether they are well motivated enough, we should gate-keep papers based on whether the method makes sense, whether it was conducted thoroughly, whether sample sizes are large enough, whether the constructs make sense, etc. If we radically raise our standards for academic rigor, we will not only prevent the floodgates from opening, but we will also ensure that papers that do get published actually produce important knowledge.
I do not mean to imply that papers today are not rigorous – far from it – but there are several instances where I se epapers published that could have larger sample sizes. Right now, they are in line with norms of the field, and their sample sizes are ‘sufficient’. But in this new crazy world, we could raise the standards and have studies incorporate more participants / scenarios / cases / etc.
As a community, if we are bombarded with potentially irrelevant papers, how will we know what papers as a community we should focus on? The fact that a paper is published is not the only way that news of a paper spreads. Papers that have more ‘impact’ and are more ‘relevant’ will get cited more, more people will talk about them, they will make their rounds on relevant social media, etc. Thus, there will still be a way for people to find the papers that they need to.
Ok. But then if you want to accept papers with lowish contribution, let these papers be published at lower-tier venues, and reserve the premium high-impact papers for the premium high-impact venues. This could be a nice middle ground approach if we lived in a world where there was no impact on a scientists’ career as to whether a paper was published in a high or low impact venue. But as long as a scientists’ career is tied to the quality of the venue they publish at, then this approach leaves many problems that are present in the status quo, most notably the incentives to lie or embellish the truth. Now if this was paired with a system whereby a scientist’s CV was not dependent on how many publications at top venues they have, then go for it! That would be lovely.
But HCI is a paper about impacting the real world, less about broad theoretical issues and the like. Thus relevance and real world impact matter more than other fields, and having papers be relevant in the future is not as big a deal. I hadn’t really thought of HCI that way, but there is something to be said about HCI being more applied than other fields, and so . It is true we are more applied than other fields, and so immediate impact matter more. Plus, the quick changing nature of tech means that the research will never truly be future proof. But regardless of this point, I think that the problems offered by the status quo outweigh the fact that HCI is a practical hands-on field.
But convincing scientists about the importance of your work is an important part of science and a skill that should be developed and reinforced. Here it’s a case of agree to disagree. I think the main value of a scientist should be to establish truth about the world, not to debate three other people that the truth you have uncovered is more important truth than some other truth. Maybe in other fields the art of convincing someone of the importance of what you are doing is needed, but I disagree.
I want to caveat, I do think that communicating your ideas to the general public is important. But that does not happen in papers. Papers are meant for scientists, and the public often doesn’t read them, both due to cultural and institutional barriers. There are other venues to reach the public (blog posts, books, interviews, talks), and there scientists should articulate clearly the contribution and importance of their work. But in those arenas.
You are just arguing for this because your papers get rejected on these grounds and you are bad at articulating your contributions. That is correct. It is also a logical fallacy (an ad hominem attack to be precise) and doesn’t really contradict my argument – the fact that I am making this argument primarily because I am a petty vengeful diva about my papers getting rejected for what I consider to be BS reasons doesn’t mean that my argument is flawed. Attack the argument, not the human. It would be like CHI establishing a process whereby to get a paper accepted you need to run a sub-4 hour marathon, and when someone complains, the CHI committee responds “Well you’re just saying that because you suck at running marathons.” Someone can be bad at running marathons, and it also be bad that paper acceptance hinges upon completing a marathon.
But I like writing lengthy discussion sections! That is amazing! I do not think we should ban discussion sections or rate a paper pooly because it has one – if writing discussion sections is your jam, keep doing it! Just don’t punish an otherwise perfect paper but who’s discussion section doesn’t convince you that this paper is the best thing since sliced bread.
Your rebuttals to the counterarguments do not convince me Alas I have neither the time nor the space to counter the counterarguments to my counterarguments to the counterarguments of my original argument. But you are allowed to disagree with me! This post is deliberately provocative. As long as I sparked some thought or conversation, then this was successful. You can always message me (see contact) if you want to email me what your counterarguments are. Or if there is someone from another field that de-emphasizes storyline / significance / contribution – I’m open to changing my mind!